Visible to Intel only — GUID: GUID-B58B596B-84D2-4073-9E95-C0CE84DB41F8
Visible to Intel only — GUID: GUID-B58B596B-84D2-4073-9E95-C0CE84DB41F8
ungqr_batch_scratchpad_size (Group Version)
Computes size of scratchpad memory required for ungqr_batch (Group Version) function. This routine belongs to the oneapi::mkl::lapack namespace.
Description
Computes the number of elements of type T the scratchpad memory to be passed to the ungqr_batch (Group Version) function must be able to hold.
API
Syntax
namespace oneapi::mkl::lapack { std::int64_t ungqr_batch_scratchpad_size(sycl::queue &queue, std::int64_t *m, std::int64_t *n, std::int64_t *k, std::int64_t *lda, std::int64_t group_count, std::int64_t *group_sizes) }
This function supports the following precisions and devices:
T |
Devices supported |
---|---|
std::complex<float> |
CPU and GPU |
std::complex<double> |
CPU and GPU |
Input Parameters
- queue
-
Device queue where calculations will be performed.
- m
-
Array of group_count parameters mg as previously supplied to geqrf_batch (Group Version) (mg ≥ 0).
- n
-
Array of group_count parameters ng as previously supplied to geqrf_batch (Group Version) (ng ≥ 0).
- k
-
Array of group_count parameters kg.
The number of elementary reflectors whose product defines the matrices Qi (0 ≤ kg≤ ng).
- lda
-
The leading dimension of Ai as previously supplied to geqrf_batch (Group Version) (ldag≥ max(1, mg)).
- group_count
-
Specifies the number of groups of parameters. Must be at least 0.
- group_sizes
-
Array of group_count integers. Array element with index g specifies the number of problems to solve for each of the groups of parameters g. So the total number of problems to solve, batch_size, is a sum of all parameter group sizes.
Exceptions
Exception |
Description |
---|---|
mkl::lapack::exception |
This exception is thrown when an incorrect argument value is supplied. You can determine the position of the incorrect argument by the info() method of the exception object. |
Return Values
The number of elements of type T the scratchpad memory to be passed to the ungqr_batch (Group Version) function must be able to hold.